Genotype Imputation and Accuracy Evaluation in Racing Quarter Horses Genotyped Using Different Commercial SNP Panels
Pereira, Guilherme L. [UNESP]
Chud, Tatiane C.S. [UNESP]
Bernardes, Priscila A. [UNESP]
Venturini, Guilherme C. [UNESP]
Chardulo, Luís A.L. [UNESP]
Curi, Rogério A. [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
The availability of large-scale SNP genotyping panels for horses has opened up new possibilities for genetic studies of the species. The first panel version contained around 54,000 SNPs (54k). The current second generation panel possesses approximately 65,000 SNPs (65k), which 19,000 of them are new markers and 45,000 were present in the first panel. The objectives of this study were to perform two-step genotype imputation between individuals in a relatively small population sample of racing Quarter Horses genotyped with the 54k or 65k panel and to evaluate the accuracy of imputation in a practical situation. Genotypes were imputed using 116 horses genotyped with the 54k SNP array and 233 animals genotyped with the 65k array. For the simulations, random samples were chosen to compose the imputed (about 50% samples) and reference populations in two scenarios. Scenario A simulated the genotype imputation in the first step (from 65k to 54k) and the scenario B in the second step (from 54k to 65k). For each scenario, the individuals genotyped with the same panel were considered as the reference population but that were included in the imputed group had their exclusive genotypes masked for subsequent evaluation of imputation accuracy. Thus, after quality control, the 54k and 65k panels contained 7,048 and 16,940 exclusive markers, respectively. The mean concordance rate was 0.9815 and 0.9751 for scenarios A and B, and the mean allelic r2 was 0.9791 and 0.974, respectively. No important influences of the genomic relationship coefficients on imputation accuracy were observed between the imputed and reference samples. Thus, two-step imputation of exclusive genotypes between the 54k and 65k equine panels resulted in a considerable increase of SNPs for all samples (12%–26%) without losses in the quality of information.
Chromosome, Concordance rate, IBD, MAF, Population structure
Journal of Equine Veterinary Science, v. 58, p. 89-96.